Closed-loop control is an effective control method to deal with uncertain disturbance in water supply canal system (WSCS). However, previous researches mainly focus on closed-loop control with a single time scale, and it is difficult to solve the uncertainty problem with a long time dimension. In this paper, a multi-scale closed-loop coupled real-time scheduling framework is established and solved by considering both large and small scale closed-loop scheduling models, so as to improve the performance of real-time scheduling in a longer time dimension under uncertain perturbations. The framework is applied to the virtual simulation test of the Xuhong River WSCS of the east route of the South-to-North Water Diversion project under different water distribution disturbances. The results show that the closed-loop control method can update the system status regularly through rolling optimization, and effectively reduce the prediction error of water level and energy consumption. In the two water distribution disturbance scenarios, the proposed method reduces the average relative deviation of water supply plan execution by 3.4 % and 5.9 %, and the unit energy consumption of cascade pumping station operation by 1.7 kWh/104m3 and 1.5 kWh/104m3, respectively, compared with the single-scale closed-loop scheduling method. In addition, the proposed method can effectively deal with the uncertainty of water distribution disturbance in limited computing time, and can meet the requirements of real-time scheduling in a long time dimension under uncertain disturbance.